U.S. DEPARTMENT OF COMMERCEyTechnology Administration

National Institute of Standards and Technology

-QC

100

.U57

1993

Semiconductor Measurement Technolog

Evolution of Silicon Material

Characterization: Lessons Learne

for Improved Manufacturin

W. Murray Bullij

The National Institute of Standards and Technology was established in 1988 by Congress to "assist

industry in the development of technology . . . needed to improve product quality, to modernize

manufacturing processes, to ensure product reliability . . . and to facilitate rapid commercialization ... of

products based on new scientific discoveries."

NIST, originally founded as the National Bureau of Standards in 1901, works to strengthen U.S.

industry's competitiveness; advance science and engineering; and improve public health, safety, and the

environment. One of the agency's basic functions is to develop, maintain, and retain custody of the national

standards of measurement, and provide the means and methods for comparing standards used in science,

engineering, manufacturing, commerce, industry, and education with the standards adopted or recognized

by the Federal Government.

As an agency of the U.S. Commerce Department's Technology Administration, NIST conducts basic

and applied research in the physical sciences and engineering and performs related services. The Institute

does generic and precompetitive work on new and advanced technologies. NIST's research facilities are

located at Gaithersburg, MD 20899, and at Boulder, CO 80303. Major technical operating units and their

principal activities are listed below. For more information contact the Public Inquiries Desk, 301-975-3058.

Technology Services• Manufacturing Technology Centers Program• Standards Services

• Technology Commercialization

• Measurement Services

• Technology Evaluation and Assessment

• Information Services

Electronics and Electrical EngineeringLaboratory• Microelectronics

• Law Enforcement Standards

• Electricity

• Semiconductor Electronics

• Electromagnetic Fields'

• Electromagnetic Technology'

Chemical Science and TechnologyLaboratory• Biotechnology

• Chemical Engineering*

• Chemical Kinetics and Thermodynamics• Inorganic Analytical Research• Organic Analytical Research

• Process Measurements• Surface and Microanalysis Science

• Thermophysics^

Physics Laboratory• Electron and Optical Physics

• Atomic Physics

• Molecular Physics

• Radiometric Physics

• Quantum Metrology

• Ionizing Radiation

• Time and Frequency'

• Quantum Physics'

Manufacturing Engineering Laboratory• Precision Engineering

• Automated Production Technology• Robot Systems

• Factory Automation• Fabrication Technology

Materials Science and EngineeringLaboratory• Intelligent Processing of Materials

• Ceramics• Materials Reliability'

• Polymers

• Metallurgy

• Reactor Radiation

Building and Fire Research Laboratory• Structures

• Building Materials

• Building Environment• Fire Science and Engineering

• Fire Measurement and Research

Computer Systems Laboratory• Information Systems Engineering

• Systems and Software Technology• Computer Security

• Systems and Network Architecture

• Advanced Systems

Computing and Applied MathematicsLaboratory• Applied and Computational Mathematics^• Statistical Engineering^

• Scientific Computing Environments^

• Computer Services^

• Computer Systems and Communications^• Information Systems

'At Boulder, CO 80303.

^Some elements at Boulder, CO 80303.

Semiconductor Measurement Technology:

Evolution of Silicon Materials

Characterization: Lessons Learned

for Improved Manufacturing

W. Murray Bullis

Materials and Metrology

1477 Enderby WaySunnyvale, CA 94087

Prepared for

Semiconductor Electronics Division

Electronics and Electrical Engineering Laboratory

National Institute of Standards and Technology

Gaithersburg, MD 20899

July 1993

U.S. DEPARTMENT OF COMMERCE, Ronald H. Brown, Secretary

NATIONAL INSTITUTE OF STANDARDS AND TECHNOLOGY, Arati Prabhakar, Director

National Institute of Standards and Technology Special Publication 400-92

Natl. Inst. Stand. Technol. Spec. Publ. 400-92, 41 pages (July 1993)

CODEN: NSPUE2

U.S. GOVERNMENT PRINTING OFFICEWASHINGTON: 1993

For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402-9325

TABLE OF CONTENTS

Page

ABSTRACT 1

INTRODUCTION 1

EARLY DEVELOPMENTS 2

BEGINNINGS OF STANDARDIZATION 3

Growth of ASTM Committee F-1 4

Some Problems 6

U.S. Government Activities 7

STANDARDIZATION OF WAFER DIMENSIONS 8

Growth of SEMI Standards 9

Internationalization 9

Other Developments 10

A MOVING TARGET 11

Silicon Standards Activities 12

Beyond Silicon Wafer Specifications 15

TODAY'S ENVIRONMENT 15

Silicon Wafer Specification Form 15

Other Developments 18

In-Process Metrology 18

Industry-Wide Planning 20

CONCLUSIONS AND RECOMMENDATIONS 21

Key Factors 21

Recommendations 23

ACKNOWLEDGMENTS 23

REFERENCES 23

APPENDIX A 27

LIST OF FIGURES

Page

1 . Flow chart showing creation of an ASTM standard test method 5

2. History of the interlaboratory precision of the measurement of silicon resistivity in the

range 0.01 to 100 fi-cm by the four-point probe method 7

3. Trends in minimum feature size, die area, and wafer diameter 11

4. History of the calibration factor for oxygen in silicon by infrared absorption 14

LIST OF TABLES

Page

1 . Parameters Most Commonly Specified for Polished Silicon Wafers 17

Semiconductor Measurement Technology:

EVOLUTION OF SILICON MATERIALS CHARACTERIZATION:LESSONS LEARNED FOR IMPROVED MANUFACTURING

Based on a paper presented at the 1992 U.S. Workshop on the Physics and Chemistry of Mercury-Cadmium-Telluride,

Danvers, Mass., October 16, 1992

with additional supplementary information

W. Murray Bullis

Materials & Metrology

Sunnyvale, CA 94087

ABSTRACT

The growth of the silicon device and integrated circuit industry has been closely coupled

with the development of materials characterization technology. This paper traces this

development from the beginning, when the industry was young and each manufacturer

had to grow its own materials, develop its processes, assemble its measurement systems

from component instruments, and fabricate its processing equipment, to the present, whena complex infrastructure supports the industry. It also describes examples of both

successful and unsuccessful developments in connection with other electronic materials.

The critical factors in the widespread standardization and application of silicon

characterization technology are: shared results and common goals among industry,

government, and universities; an efficient mechanism for development of consensus

standards with adequate expertise; standardization of parameters needed for orderly flow

of materials in manufacturing together with standardized terminology, test methods, and

formats to support detailed purchase specifications; and refinement of the measurement

technology as the industry develops and becomes more sophisticated. Finally, a good

methodology for process diagnosis and control is an essential feature for applying test

procedures to the manufacturing environment. The successes and failures which occurred

during the course of developing metrology tools for silicon can provide some insights into

similar developments for the special needs of other electronic materials. In addition,

many of the test procedures and standards developed for silicon materials and supporting

technologies can be applied directly.

Key words: ASTM; characterization techniques; materials characterization; NIST; SEMI;

silicon; standards; wafer dimensions

INTRODUCTION

The growth of the silicon device and integrated circuit industry has from its inception been closely

coupled with the development of characterization technology for silicon materials. We have to recognize

that silicon enjoys a major advantage over compound semiconductors in that the single element matrix

provides considerable simplification in both preparation and analysis. The material is an excellent vehicle

for probing the inner workings of the solid state as well as being abundant, relatively inexpensive, and

the backbone of a major industry. Nevertheless, the lessons learned during the long evolution of

— 1 —

metrology for the silicon industry may help in defining efficient paths for metrology development in

connection with other, more specialized electronic materials.

However, there is also a nontechnical factor which must be taken into account when trying to apply the

experience gained in silicon to other materials. Relative to the various electronic materials, electronic-

grade silicon enjoys a very large market, and thus the industry is able to support a substantial

infrastructure. Nevertheless, the size of the electronic-grade silicon market is tiny when considered in

terms of markets for commodity materials such as steel, aluminum, chemicals, and the like. In fact,

much more silicon — of course, of metallurgical grade — is used in the aluminum and steel industries

than is used for semiconductor devices and integrated circuits.

The requirements for purity, uniformity, and dimensional control of electronic-grade silicon are severe

and have from the beginning taxed the metrological and contamination control capabilities of supporting

industries. Because of the relatively small volumes involved, it has proved to be quite difficult to

encourage these supporting industries to develop their capabilities to the degree necessary for advancing

silicon technology. Materials with even smaller markets would be expected to encounter similar

difficulties, but may be able to utilize much of the infrastructure established for the silicon industry.

This paper is not intended to be a complete history of the development of silicon material metrology and

standards. Such a history would require a rather large volume to include all the myriad of significant

contributions to the field. Rather, it focuses on issues which show the impact of metrology development

and standardization efforts on the growth of the silicon industry. We first review the early history of

silicon materials characterization. At this time, the industrial base was highly customized and each major

manufacturer of devices made its own equipment and grew its own materials. Then we consider the

beginnings of test procedure standardization about 30 years ago. This led to standardization of the

dimensional properties of silicon wafers in the mid-seventies which enabled the development of the

merchant equipment industry which now supports the silicon device industry.

As time went on, the precision and sensitivity requirements of various test procedures increased and newcharacteristics had to be determined. This trend continues today as metrology development evolves to

meet the requirements of advanced sub-micrometer ultra-large scale integrated (ULSI) circuit technology;

however, the environment for these developments has changed markedly from that which existed in the

early days of the industry. Finally, we identify the critical factors in the widespread application and

standardization of silicon characterization technology and draw some conclusions regarding extensions

to the development of metrology for other materials.

EARLY DEVELOPMENTS

In the decade following the invention of the transistor in the late forties, there was a period of intense

materials characterization which laid the foundations for our current knowledge of the electronic, optical,

and mechanical properties of silicon. In particular, the group at the Bell Telephone Laboratories

published prolifically during the early fifties — journal articles, technical papers, and books — on all

aspects of semiconductor technology from basic theory to practical application, from measurement

techniques to process technology. In addition, a variety of groups from other major companies,

universities, and government laboratories made important contributions to this effort. During this period,

the fundamental aspects of semiconductor behavior were established firmly, and basic metrology tools

were demonstrated. It was a very exciting time as new ideas and theories were subjected to experimental

scrutiny.

— 2 —

The great advances made during this period unfortunately led to a degree of complacency regarding the

completeness of the state of knowledge of silicon technology by 1960. Nevertheless, the importance of

the widespread dissemination of the basic technologies cannot be overemphasized. The flow of

information throughout the young industry fueled rapid advances in both technology and applications and

set the stage for future development and growth.

The concept of doping (which had been first articulated in connection with semimetals in the U.K. during

the thirties) and the one-to-one correspondence between the dopant atom density and the free carrier

density were experimentally demonstrated. Minority carrier transport properties and p-n junction

characteristics were modeled and controlled. Photoconductivity and optical absorption characteristics

were demonstrated. Surface states were analyzed and basic models established. Both modified

Czochralski and floating zone techniques were developed for growing "large" single crystals, and the

process for growing dislocation-free silicon crystals (now routinely expected when silicon is mentioned)

was developed and introduced into production. Chemical vapor deposition techniques for producing

silicon films on silicon and other substrates were introduced. Theoretical and experimental band structure

investigations led to an understanding of transport phenomena in both silicon and germanium.

On the metrology side, the four-point probe was rediscovered and applied to the measurement of

resistivity of semiconductor crystals and slices. Extensive application of the Hall effect allowed

determination of carrier densities and mobilities over a wide temperature range leading to basic

knowledge about impurity energy levels in the forbidden band gap. Photoconductivity decay was

introduced as a direct measure of minority carrier recombination lifetime. The infrared absorption band

at a wavelength of about 9 /xm was used to determine the oxygen content of silicon crystals. Both x ray

diffraction and optical techniques were used for orienting wafer surfaces. The thermoelectric effect and

point-contact rectification were applied to the determination of conductivity type. Various preferential

chemical etchants were developed to expose dislocations and crystal defects which intersected wafer

surfaces.

During the initial phase of these developments, interest was focused on germanium materials and devices.

Because of its relatively low melting point, this material was easy to prepare and process. However,

because of its narrow band gap, devices could be operated only at room temperature; even the heat

generated in a car parked in the sun was sufficient to cause germanium transistors to become inoperative.

Hence, in the mid-fifties, the major emphasis shifted from germanium devices to silicon devices. Anelemental semiconductor like germanium, silicon was much more abundant. In fact, silicon is the second

most abundant element on Earth. Not only did silicon permit the operation of devices at higher

temperatures because of its wider band gap and good thermal conduction characteristics, but also it had

the inherent strength to withstand handling during manufacture and use. In addition, it was possible to

passivate the surface with thermal oxide. Although this characteristic was soon to have far reaching

implications with regard to process standardization and the development of integrated circuits, the device

industry at this time remained in its custom mode with each major manufacturer designing and building

its own equipment, growing its own materials, and developing its own process technology. Device

technologies evolved rapidly from grown junction to alloy junction to diffused junction, from mesa to

planar. Standardization was yet to come.

BEGINNINGS OF STANDARDIZATION

By the early sixties, the two basic concepts which have fueled the growth of the microelectronics age

were established. The viability and advantages of planar technology for device fabrication and the

— 3 —

feasibility of placing multiple components on a single chip to form an integrated circuit were both

demonstrated. But all was not completely rosy. A fledgling merchant materials industry was growing

up and attempting to supply silicon wafers to device manufacturers. Measurement conditions and

procedures were imprecisely defined and there were many disagreements over material properties as

measured by supplier and user. This is a classic example of the fact that the demands on definitions and

metrology made by the manufacturing environment are much more severe than those made by the

research environment which seeks to demonstrate a scientific principle rather than to replicate nominally

identical products [1].

These disagreements at the supplier-user interface demonstrated a need to standardize the applicable

measurement techniques. This need was met through the American Society for Testing and Materials

(ASTM) Committee F-1 on Electronics. In 1955, this Committee had evolved as the first "end-use"

committee in ASTM from a subcommittee on non-ferrous metals for vacuum tube application. Its scope

covered development of standards for materials for all types of electron devices. By 1960, through the

visionary leadership of its early chairmen, Sid Standing of Raytheon and Frank Biondi of Bell Telephone

Laboratories, the Committee had made a rapid transition to cover semiconductors [2].

Growth OF ASTM Committee F-1

Over the next decade, standardized methods were developed for the basic parameters by which silicon

is specified. These included all the techniques reported during the fifties: the four-point probe method

for measuring resistivity and radial resistivity variation; the photoconductivity decay method for minority

carrier recombination lifetime (also developed by the Institute of Electrical and Electronic Engineers -

IEEE); x-ray and optical methods for crystallographic orientation; chemical etching to reveal dislocations,

stacking faults, and other crystallographic defects; various methods for conductivity type determination;

Hall effect for carrier density and mobility; and infrared methods for the determination of oxygen content

and epitaxial layer thickness.

These methods were the result of intensive efforts at major industrial laboratories, particularly Bell Labs,

IBM, RCA, and GE, and at the National Bureau of Standards.^ Not only did these organizations

provide the research necessary to develop the methods to a state suitable for use in production

environments, but they also provided the leaders for the standardization activities. Without this intensive

refinement of measurement methods during the sixties, the pace of development of metrology for the

industry would not have been so rapid.

The successful development of test methodology for silicon materials was due in large part to the rigorous

procedure by which consensus was reached and the efficacy of the method demonstrated [3] as outlined

in the flow chart in Figure 1 . Two blocks of this flow chart are emphasized. The first indicates the

people required to carry out the activity; we will consider the effect of this factor later. The second

identifies the interlaboratory evaluation of test methods. Despite the fact that this rigorous process was

indeed time-consuming, the results afforded significant benefits to the participants. In particular, these

cooperative experiments allowed participating organizations to determine their abilities to carry out

important measurements and frequently highlighted difficulties in one location or another which could

then be corrected using the new information derived from the experiment. One frequently heard

^ The National Bureau of Standards was renamed the National Institute of Standards and Technology

(NIST) in 1988.

— 4 —

Figure 1 . Flow chart showing creation of an ASTM standard test method. (Fronn Ref. 3.)

complaint about this procedure is that it is unduly time-consuming. Experience over the years has

indicated that if there is a real, well-understood need for the test method, the entire procedure from

— 5 —

conception to publication, including completion of preliminary testing, can be completed in less than

2 years. This is, however, not the norm; average development time is 3 to 5 years. Nevertheless, a

more rapid and effective procedure has yet to be demonstrated. Efforts to speed up the process by short-

circuiting important elements of the developmental process have frequently led to additional delays.

There was also intensive development of methods to control contamination levels in air, chemicals, and

other fluids, and on surfaces. A subcommittee on this topic was organized at the time Committee F-1

was established. It conducted a series of landmark symposia which defined the industry requirements for

contamination control and led to the development of nearly 20 standards. Many of these early ASTMTest Methods were incorporated into the initial Federal Standard on Clean Rooms [4]. This standard has

evolved over the years and several updated editions have been published; it is now being further revised

to accommodate the more rigorous contamination control requirements associated with ULSI technologies.

Some Problems

At the same time that the Committee was having considerable success with developing standards for

silicon substrates and contamination control, it failed to duplicate these successes in several other areas.

There were two notable examples of this. One instructive example of F-1 failure occurred in the area

of trace analysis. It was broadly recognized that trace analysis was a critical aspect of semiconductor

technology because the properties of the materials are controlled by minute amounts of foreign atoms —some of which are intentionally introduced and some of which are not. The Committee attempted to

establish a group to refine chemical and physical methods so that they could achieve the sensitivities

required. In this case, it proved impossible to attract analytical specialists to participate actively in a

group with highly specialized and very demanding goals. Some related activities were subsequently

conducted in other ASTM committees, notably Committee E-42 on Surface Analysis, but standardization

of these analytical methods in the regimes appropriate to silicon material and device technologies has

lagged to this day.

The second example occurred in the area of standards for testing phosphor materials used for cameras

and displays. Although there was general agreement that such standards would be beneficial to the

industry, the effort to develop them failed because common ground could not be established. The

procedures for preparing the phosphors were considered to be highly proprietary; since different

procedures required different properties in the starting materials, even discussing starting material

properties was considered to be off-limits because they might reveal some aspect of the subsequent

processing. Further, there was the feeling in some quarters that the properties of the starting materials

were not so very important after all because of the changes made during processing. Thus, no useful

dialogue between suppliers and users of the materials could be established and the effort did not achieve

any results.

Since the mid-sixties, we have learned a great deal about establishing common ground for standards, and

it is possible that if the effort had been conducted with the benefit of this knowledge, we might have made

more progress. Among the issues which can be discussed at the outset are definitions of commonly used

terms. Although good definitions reduce the ability of suppliers to play "specsmanship" games, there

is usually enough interest in commonality in this area to get some useful dialogue begun. The second

area is in test methodology for very basic properties such as composition or geometry. Once rapport is

established in these areas, it is sometimes possible to extend the effort to more controversial areas. This

approach is now being used with modest success in connection with testing for magnetic disk

technologies.

— 6 —

100

0.1

'INDUSTRIAL ROUND ROBIN PRIOR TO

AN A.S.TM STANDARD

ROUND ROBIN TEST

OF A.S.T.M. STANDARD

ADOPTED 1964

VALUE DESIRED FOR MATERIALS

EXCHANGE - 1%

J L 1 J L

58 60 62 64YEAR

66 68

Figure 2. History of the interlaboratory precision of the measurement of silicon resistivity in the range

0.01 to 100 Q cm by the four-point probe method. (The vertical bars represent the range in samplestandard deviation obtained in the interlaboratory (round robin) tests.) (From Ref. 5.)

U.S. Government Activities

Early in the history of ASTM Committee F-1, a relationship was established with the National Bureau

of Standards. Around 1960, the Bureau's Tube Laboratory began a shift toward semiconductor

technology. This grew in concert with the growth of ASTM Committee F-1, and in the late sixties, the

semiconductor effort was formalized into a major program under the direction of Charles Marsden and

Judson French. This program was largely supported by Defense Department, Energy Department, and

NASA funds. One of the major accomplishments of the early phase of this program was a significant

improvement in the precision of four-point probe resistivity measurements, as shown in Figure 2. This

— 7 —

effort is a classic example of measurement method development. It involved not only considerable

refinement of the measurement conditions but also a great deal of industrial cooperation. The results

were validated by interlaboratory testing in Committee F-1 [5].

During this same period, two highly significant developments occurred within the military and space

agencies of the U.S. Government. First, Joe Brauer's group at Rome Air Development Center developed

a revolutionary procedure for qualification of integrated circuits for military application. When it was

first introduced in 1968, MIL-STD-883 was considered by many to be too difficult and unnecessary;

many vendors sought and obtained waivers from its rigid testing procedures. Its most obvious limitation

was that it continued the philosophy of "testing-in" rather than "building-in" quality. Nevertheless, the

discipline imposed by the testing hierarchy, particularly when combined with efforts to trace causes of

testing failures, provided a methodology which could be applied to process control and improvement.

Although many in the United States failed to take advantage of this opportunity, it was not lost on the

young, but rapidly developing, Japanese semiconductor industry. Within a few years following its

introduction, the principles of MIL-STD-883 were being strictly applied to good advantage in Japan.

Coupled with the principles of quality through statistical process control as espoused by Deming and

others, MIL-STD-883 became a formidable weapon in the efforts of the Japanese to improve both quality

and manufacturing efficiency.

The NASA approach was different from, but complementary to, the more traditional approach taken by

the Air Force. In the late sixties, Dr. A.M. Halliday and his group at the Marshall Space Flight Center

established a program known as the Line Certification Program, specifically directed toward integrated

circuit fabrication. This program focussed on control of the incoming materials (such as silicon,

chemicals, gases, and parts), the clean room environment, and fabrication processes (such as oxidation,

patterning, ion implantation, diffusion, metallization, and passivation). Although far from perfect, this

program, when applied, provided process control features similar in many ways to those found in the

Deming approach. Significant improvements in both quality and manufacturing efficiency could be

realized when the techniques of this program were applied, but frequently they were viewed as too costly

and complex to be practical.

STANDARDIZATION OF WAFER DIMENSIONS

In the early seventies, the industry was expanding rapidly and silicon material was in short supply. At

this time, wafers were being manufactured with a great variety of geometrical characteristics to respond

to the myriad of requirements of the various device lines. This required vast inventories of materials of

similar, but not interchangeable, geometries.

Initial standardization efforts were undertaken on an ad hoc basis by silicon suppliers [6]. A limited

number of values for wafer diameter, flat length, thickness, and thickness variation was proposed in order

to reduce the proliferation of wafer sizes and improve manufacturing efficiency. It should be noted that

although a system of flats on the periphery of the wafer was developed to differentiate wafer conductivity

type and surface orientation, values of electrical and crystallographic properties, which cover a wide

range and are generally specific to a particular device process were not included in the specifications.

In fact, this concept, the standardization of geometrical properties while allowing flexibility in application-

specific material characteristics, is a key element in the successful industrial application of the silicon

metrology development and standardization activities.

— 8 —

Growth of SEMI Stamdards

These wafer specifications were formalized by the Semiconductor Equipment and Materials Institute

(SEMI)^ in 1973 [7] and within a few years became generally accepted throughout the free world [8].

The SEMI specifications were not only successful in helping to alleviate the material supply crisis which

had precipitated their development; they also had several other significant impacts.

First, they were instrumental in increasing industry awareness of the ASTM Test Methods which were

continuing to be produced by Committee F-1. There was a general awareness that standardized test

procedures were necessary to support the specifications. By the end of the seventies the Committee had

developed test methods for the dimensional properties covered by the SEMI specifications as well as manyof the other properties which had to be specified in individual purchase documents; all applicable test

methods were referenced in the SEMI silicon wafer specifications.

Second, although the SEMI specifications were introduced at about the time the industry was shifting to

76.2-mm (3-in.) diameter wafers, they were extended periodically to cover larger diameter wafers and

so facilitated the orderly production of 100-, 125-, and 150-mm wafers as the industry evolved. The

concept of dimensional standardization became widely accepted and expected as the specifications for

these larger diameters were introduced.

Third, the existence of standardized geometries for wafers made it possible to design and construct wafer

handlers, transport systems, and processing equipment on an interchangeable basis. This, together with

the almost universal adoption of the planar process for fabricating devices and circuits, allowed the

establishment and orderly growth of a merchant semiconductor process equipment industry. No longer

did each device manufacturer have to design and build its own processing equipment on a custom basis;

it is now possible to obtain such equipment commercially. In fact, merchant equipment suppliers nowprovide a great deal of the process development activity in the industry, frequently in partnership with

Semiconductor Manufacturing Technology (SEMATECH) or a major device house.

Internationalization

Fourth, the worldwide acceptance of the SEMI standards for silicon wafers accelerated the

internationalization of the standardization efforts [9]. ASTM Committee F-1 had established informal

liaison with Committee FNM A9 (now NMP 221) of the Deutsches Insititut fiir Normung (DIN) in

Germany in the late sixties. This cooperation grew and was formalized in 1976. During this period,

technical differences between ASTM and DIN test methods for silicon characterization were virtually

eliminated so that test methods generated by either organization are cited in and can be used in support

of the SEMI silicon wafer specifications.

Late in the seventies, initial contacts were made with the Japan Electronic Industry Development

Association (JEIDA) which was developing standard test methods for characterizing silicon wafers in

Japan. These contacts developed into an active partnership in the early eighties. Although significant

differences exist in geometrical characteristics of silicon wafers in Japan and in the United States,

cooperative activities are continuing in an effort to reduce these differences.

^ The Semiconductor Equipment and Materials Instimte was renamed the Semiconductor Equipment

and Materials International in 1987 to reflect the global nature of the industry.

— 9 —

Other Developments

During this period, the NBS program on semiconductor measurement technology was expanded

significantly with major funding from the Defense Advanced Research Projects Agency (ARPA) and the

Defense Nuclear Agency (DNA). In-house work on wire bonding, microelectronic test structures, and

linewidth measurements responded to a number of critical needs. Under the ARPA program, more than

20 contracts were awarded to industry, academic, and government laboratories over a six-year period to

support the development of a wide range of innovative measurement techniques. Topics covered included

development of micrometer-scale imaging for acoustic microscopy, detailed investigation of the Si-Si02

interface, study of the transfer of metal ions from solutions to wafer surfaces, surface quality tests for

sapphire substrates, development of specialized microelectronic test structures, and determination of

electron and hole mobilities in silicon as a function of dopant density and temperature. Some of these

were immediately successful and were applied directly in the industry; others were somewhat ahead of

their time and did not meet with wide acceptance [10]. Overall, however, this program provided a very

useful boost to the development of measurement technology for state-of-the-art integrated circuit

fabrication.

NBS also introduced Standard Reference Materials (SRMs®^) specifically for the semiconductor industry

during the early seventies. The first of these was a set of two wafers certified for resistivity as measured

by the standard four-point probe method [11]. Other resistivity SRMs were issued later in the decade.

This program was expanded during the eighties to include SRMs for linewidth measurement on

photomasks, spreading resistance calibration, and oxide film thickness. A key point is that, in most

cases, the certified value is based on measurements made in accordance with a consensus procedure

developed and tested by interlaboratory experiment in ASTM Committee F-1. Despite the fact that

absolute standards for these measurements do not exist, and so the bias of the certified value cannot be

estimated, these SRMs provide a common basis for calibration and control of test equipment in the

industry and thus facilitate supplier-user exchange.

At the same time, ASTM Committee F-1 extended its activities into other areas. Building on the work

in the NBS program, a number of ASTM specifications and test methods were developed for

interconnection bonding wire and bond testing. Later the work on bonding broadened to include tape-

automated bonding, a current major activity of the Committee. A second successful area of activity was

in radiation hardness, also undertaken with the initial assistance of DNA and NBS. The cornerstone of

this activity is cooperation and coordination with many other governmental and industrial groups

concerned with device quality and radiation hardness assurance. The subcommittee has developed more

than 30 standard test methods, practices, and guides and continues to produce several new and revised

standards each year.

Groups addressing gallium arsenide, gadolinium gallium garnet, and passive microwave devices were also

established in Committee F-1. Of these, only the gallium arsenide group now remains active. However,

this group has developed only a few standards. There are two reasons for this. First, many of the

relevant test procedures, especially those for dimensional properties, are the same for gallium arsenide

^ The term "SRM" is a Federally registered trademark of the National Institute of Standards and

Technology and the U.S. Government.

— 10 —

1960 1970 1980 1990 2000 2010

Year

Figure 3. Trends in minimum feature size, die area, and wafer diameter. (Data shown are from Ref.

12.)

as for silicon; consequently they did not have to be reinvented. Second, the dominant commercial activity

in III-V compound semiconductors is in Japan rather than in the United States; thus the U.S. group is

hampered by sub-critical size.

A MOVING TARGET

Throughout the eighties, silicon technology continued to advance. Dynamic random access memory(DRAM) density quadrupled roughly every 3.2 years. To support this increase in circuit density, not

only were new design and process technologies developed but also the reductions in feature sizes and

increases of die area and wafer diameter continued to progress along their historical exponential trends:

minimum feature size was halved every 6.4 years, die area doubled every 5.4 years, and wafer diameter

doubled every 9.5 years. These trends, illustrated in Figure 3 together with historical data and future

projections [12], placed more severe requirements on silicon materials, particularly in regard to control

of bulk defect properties, surface particulate and metal contamination, and wafer flatness. Precision and

sensitivity requirements of traditional measurements steadily increased, and new characteristics had to be

measured.

— 11 —

Early in the eighties, U.S. device manufacturers began to understand that meaningful two-waycommunication and cooperation with their silicon material and equipment suppliers is essential for

efficient production of quality parts. Whereas once the material, equipment, and device operations had

all been part of the same organization, the rise of the merchant material and equipment industries tended

to build barriers between the various functions. It is instructive to note that this lesson was learned

earlier in Japan than in the United States, contributing to the earlier Japanese success in building high-

quality semiconductor devices and circuits. Major U.S. device companies began cooperative programs

with their suppliers, which were extremely effective for their participants in advancing the performance,

quality, and consistency of silicon materials. Also during this period, the Semiconductor Research

Corporation (SRC) and later SEMATECH fostered cooperative interactions between various portions of

the industry on a somewhat wider scale. SRC focused on university research programs while

SEMATECH emphasized unit processes and process equipment development.

Silicon Standards Activities

To respond to the new demands being made of silicon wafers, both SEMI and ASTM expanded their

standards development activities. SEMI developed specifications for silicon and silicon-on-sapphire

epitaxial wafers, solar-grade silicon, silicon test wafers, and polycrystalline silicon. The specifications

for many of the dimensional tolerances on silicon wafers were tightened, and a specification was

developed for 200-mm wafers with a notch fiducial instead of a flat. Specifications for gallium arsenide

wafers for both optoelectronic and integrated circuit applications were developed and published. These

followed the same pattern as the silicon wafer specifications, with standard values for dimensional

parameters and flexibility for electrical and crystallographic parameters. It is also instructive to note that

the standard test procedures cited in support of these gallium arsenide specifications were all initially

developed for silicon.

New ways of measuring many parameters were introduced by a number of relatively small instrument

houses. In general, these involved computer-controlled instruments, frequently with positioning and

scanning capabilities. Although the repeatability of such instruments was usually better than that of the

previously available techniques, comparisons between measurements made with different instruments were

often difficult because of differences in their construction or algorithms, not all of which are readily

apparent to the user. Thus the instrument behaves as a sealed "black box" with stipulated inputs and

automatic outputs. Nevertheless, standardization of test methods has continued into this environment,

even though frequently the instruments are not sufficiently defined. There are three possible choices for

standardization under such conditions:

1. Generate a referee method based on a complete description of the instrument and the

associated procedure which is usually based on manual data acquisition and analysis. Use

this referee method to generate calibration and verification standards for use with the black-

box technique.

2. Describe the procedure in terms of the data to be collected and the mathematical relationships

to be used for analysis, allowing variety in the actual implementation of the algorithms. If

desired, instrument verification procedures may be included to evaluate the suitability of the

instrument, and standard sample data sets can be provided to test the software routines.

: > / 3. Describe only a skeleton of the method, relying on the manufacturer's instructions for the

actual procedure. Such a method provides little real information; however, it is also possible

to include instrument verification routines and standard sample data sets as in case 2.

— 12 —

All of these procedures have been employed in various silicon related test methods standardized byASTM. Case 2 is generally the preferred one because it provides the greatest amount of information to

the user and facilitates valid comparisons between competing systems while allowing the instrument

designer maximum flexibility in implementing the method. Nevertheless, there are circumstances in

which one of the other cases may be all that can be agreed upon. Sometimes, agreement cannot be

reached at all.

An example of a case in which it has not yet been possible to determine an acceptable standardization

procedure for a test based on black-box instrumentation can be found in the area of wafer inspection.

Inspection of silicon wafers for particles and other surface contamination has long been a key quality

control tool. For many years, this has been accomplished by means of visual inspection under controlled

illumination in a darkened environment. As critical particle sizes and counts became smaller, automated

particle detection systems were rapidly introduced into research and production. However, development

of standards for these systems proved to be an elusive task because these instruments are essentially

computer-controlled black boxes, the detailed workings of which are generally unknown to the user.

Differences in design from one system to another result in differences in results obtained. Further, the

physics of the process is both extremely complex and incompletely understood. Although test artifacts

based on latex spheres or lithographed patterns are widely used, the relationship between these surrogates

and "real" particles is also not well established. Consequently, there is no "right" way to perform the

determination, and so there is no technical basis for a "standard" procedure. The silicon standards

community is still wrestling with this issue, both in SEMI and ASTM.

Improperly executed and continuously revised standards can confuse as well as clarify. Early in the

eighties, there was renewed interest in more precise control of the oxygen content in silicon wafers

because of the influence of the oxygen on both the gettering capabilities and mechanical strength. Upto this point, there had been considerable confusion in both the industry and research communities

because of the adoption of conflicting values for the calibration factor relating the 9-/Lcm absorption

coefficient to the oxygen concentration. ASTM had adopted three different values between 1964 and

1980; the last of these was the one which had been adopted by DIN. Still a fourth value was adopted

by JEIDA in the early eighties. The largest of these four factors was nearly twice the smallest. With

the resurgence of interest in this topic, many different individual research studies were carried out; these

led to additional values for this factor which tended to converge in the neighborhood of 6 ppma/cm~^

These results are summarized in Figure 4. To resolve the issue, an extensive international experiment

was conducted jointly by ASTM, JEIDA, SEMI, DIN, and Academia Sinica, with the cooperation and

participation of numerous government and industrial organizations around the world [13]. This

experiment brought harmony to the community since all major standards bodies have adopted its result.

Certified reference materials for oxygen determinations referenced to this calibration factor are available

from both JEIDA in Japan and the Community Bureau of Reference (BCR) in Europe. Soon, they will

also be available in the United States from NIST. Although the final result appears to be the most nearly

correct, the circuitous path which led to it was most undesirable. Some revision of standard test methods

and specifications is essential to maintain technical currency and to strive for the best possible solutions,

but too frequent change is very unsettling to the user community. Thus timing is seen to be a critical

factor in issuance of standards. If they are issued too early, the chances of selecting incorrect values and

inappropriate procedures are greater, but if they are delayed too long, there are too many different

choices already available in the marketplace and standardization becomes difficult. Even if

standardization occurs evenmally in the latter case, the long delay may cause loss of some of the

economic benefits.

— 13 —

Eo

Eaa

0+->

o

LL

co

03

n

U

14

12 -

10 -

8 -

5 -

4

in^cO^NNNNODaDOOCOOOCOOO0^0^0)0)0)0^(y)(y)(y)0)(y)(y)(y\ a\(yi^ V- T— T— T— A— T— 1— -r-T— T-T— T— V-

Year Reported

Figure 4. History of the calibration factor for oxygen in silicon by infrared absorption. Values shownas + were adopted by various standards organizations as follows: A: "Kaiser & Keck" used in original

ASTM test nnethod (F 45) from 1 964 to 1 970; B: "Old ASTM" used in ASTM test method (F 1 21 ) from

1 970 to 1 980; C: "DIN" or "New ASTM" used in DIN test method (50438/1 ) from 1 974 and in ASTMtest method (F 121) from 1980; D: "Guo Biao" used in Guo Biao test method (Peoples Republic of

China) from early 1980s; E: "JEIDA" value used widely in Japan from early 1980s; F: "IOC-88" value

being adopted by all standards organizations. (Data are taken from Table III of Ref. 13; see this

reference for the origins of values and relevant citations.)

Toward the end of the eighties, another major trend became evident in the semiconductor industry. There

was a resurgence of interest in improved quality which led to the remtroduction of statistical process

control procedures and related quality systems in U.S. industry. Responding to this change, too, has

proved to be a challenge for the silicon standards community. A small, but very important, initial step

was taken when SEMI shifted the silicon wafer specifications from "goal post" type upper and lower

specification limits to a target value with tolerance. From this point it is not too great a transition to go

to true distributional specifications — in-process statistical distributions based on process or product

measurements over time— with the goal of continuously reducing the variability of the process. Another

factor in the quest for ever-decreasing variability in manufacturing processes is the increased need for

more precise in-line control measurements, many of which do not even exist. In addition, although there

has been great progress in development of process models, these are still in a rudimentary stage and need

further refinement.

— 14 —

Beyond Silicon Wafer Specifications

The SEMI standards program also broadened its scope during the eighties in response to the needs of the

industry. It formally became an international organization with active standards development groups in

Japan and Europe as well as in the United States. Whereas in the early days the silicon materials activity

had driven the program, the growth of the equipment industry led to a major shift in emphasis.

Equipment automation became a dominant factor in driving the standards program in response to issues

such as wafer transport, machine compatibility, communications, and contamination control. This wasa logical direction for the program to take because wafer standardization alone, which had opened up the

possibilities, could not drive the development beyond a certain point.

SEMI also began standards development activities in other areas such as pure chemicals and gases,

facilities, safety, and packaging to serve the broader needs of the industry. These activities provide

standards for the infrastrucmre which is essential for efficient manufacturing of semiconductor devices

and integrated circuits. Many of these standards can be expected to be directly applicable to the

production of electronic materials and devices other than silicon.

TODAY'S ENVIRONMENT

The environment in which the challenges of silicon metrology development are being met has changed

significantly from that of the early days. The resources available for development of metrology for

silicon today are significantly fewer than those which had been available before. Most of the large user

companies no longer support extensive metrology research and development and government programs

on silicon materials technology are very limited. Instead, the main driving forces for development of

standardized test methods are coming from both silicon suppliers themselves and relatively small

measurement equipment manufacturers, each of which wants to get its "better" characterization tool to

market.

Nevertheless, the demands on the metrology for silicon wafers continue to be very great. These have

been clearly documented in several recently issued reports based on a number of sources: a JEIDA study

[14], a workshop co-sponsored by SEMATECH, ASTM, SEMI, and NIST [15], a SEMI/JEIDA joint

technical conference [16], and a sUidy at the National Institute for Standards and Technology [17]. In

response to this need, SEMI, ASTM Committee F-1, and NIST all are continuing active development of

metrology and standards.

In the past few years, the SEMI Silicon Wafer and Silicon Epitaxial Wafer Standards Committees have

developed several new specifications. These include a specification for a defined wafer coordinate

system, a specification for assigning addresses to specific points on an unpatterned silicon wafer, a

specification for 100-mm-diameter dielectrically isolated wafers, and a specification for 300-mm-diameter

wafers to assist in guiding the development of equipment and processes which will be needed in the latter

half of this decade.

Silicon Wafer Specification Form

One of the most significant recent SEMI developments is a standard format for silicon wafer

specifications [18] which collects all the parameters which might be specified and provides a common

— 15 —

vocabulary for commercial transactions. This document is a very useful adjunct to the conventional wafer

specifications [7] and is intended to assist purchasing agents and specification writers in understanding

the complex world of silicon specifications. It was first published in 1990 after several years of intensive

development with inputs from both Japan and the United States. The 1993 edition of the standard is

reproduced as Appendix A. The standard is in three parts: a brief introductory text; the form itself; and

a supporting table of references. The introductory text provides some guidance as to how to use the

form. It is intended to be used in conjunction with the SEMI polished wafer [7] or epitaxial wafer [19]

specifications which provide complete references to acceptable test methodology, standardized definitions

of parameters (such as flatness, resistivity, bow, warp, sori, etc.), and standard values for manyparameters. The table of references provides information as to whether (and in which standard or

standards) the value of the parameter is standardized, location (if any) of a definition or description of

the item, and standard test methods for the parameter as developed by ASTM, DIN, or JEIDA. Note

that some JEIDA test methods have been issued as Japan Industrial Standards (JIS) while others have not.

In general, the test methods listed are technically equivalent, although in some cases there are choices

to be made between methods which may yield different results. It is always important when specifying

a parameter to specify the test procedure by which its value is to be verified.

The form itself is in four sections. The first contains general information about the order to identify the

particular material being described. The second contains parameters appropriate to polished wafers,

including epitaxial substrates. The third contains parameters necessary for specification of epitaxial

layers. The fourth and newest section contains parameters appropriate for specification of epitaxial

wafers with buried layers. Although most buried layer work in the United States is performed by the

device manufacturer, the material suppliers are expected to provide such products in Japan. This mayindicate the emergence of another trend: device manufacturers in Japan are beginning to expect their

material suppliers to perform more of the front end processing, even including some initial patterning

steps.

The specification items are grouped according to the type of characteristic. The following groups are

identified for polished wafers: general, electrical, chemical, structural, special features (added during

wafer manufacture), mechanical, and front and back surface visual inspection. There is a place to insert

the specified value for each item or to cite a standardized value. In addition, there are columns to

indicate the level of testing required and whether the testing is to be carried out on the product wafer or

on a test piece.

Some items are identified as being required when ordering any silicon wafer; these include crystal growth

method, crystal orientation, dopant, conductivity type, resistivity, and certain mechanical dimensions,

including diameter, flat or notch dimensions and locations, thickness, surface orientation, and edge profile

contour. These represent the minimum specification list for obtaining any type or quantity of silicon

wafers. Something must be entered for each of these items, even though any particular item (such as

edge contour or resistivity value) may not be important for the intended application. In such cases, either

a very wide range (e.g., 0.1 to 99 fl cm for resistivity) or "any" may be entered.

Parameters which are most commonly specified are listed in Table 1. Those which are required are

indicated by superscript diamonds (*). One of the purposes of the format is to provide a standard

method for indicating specifications for other parameters which are not yet so well understood and widely

utilized. These include minority carrier lifetime, metal contamination level, surface organics, swirl,

shallow pits, oxide precipitation, roughness and brightness of the back surface, and special features added

in wafer manufacture such as wafer identification marking, extrinsic gettering, backseal characteristics,

and the like. In many cases, the test methods for these parameters are not standardized; in some cases,

— 16 —

Table 1 . Parameters Most Commonly Specified for Polished Silicon Wafers

Group ParameterStandardized

Value

General Growth Method^ and Crystal Orientation^ NoConductivity Type and Dopant NoNominal Edge Exclusion for Fixed Quality Area No

Electrical Resistivity^ and Radial Resistivity Variation No

Chemical Oxygen and Radial Oxygen Variation NoCarbon No

Structural Dislocation Etch Pit Density NoSlip, Lineage, Twin NoOxidation Induced Stacking Faults No

opeciai reatures

Mechanical Diameter,* Thickness,* and Thickness Variation Yes

rial or Noicn uimensions ana Locaxions T es

Edge Profile* Yes

Surface Orientation* Yes

Shape: Bow, Warp, or Son Yes

Flatness No

Visual Inspection Front Surface Yes

Back Surface Yes

These items must be specified when ordering silicon wafers.

adequate test methods do not even exist. An example is the case of surface metal contamination levels

where the desired specification limit is about the same as the sensitivity of the test method technology.

To describe epitaxial wafers, additional parameters must be specified. Required parameters include wafer

description (conductivity type of layer and substrate) together with the dopant, net carrier density, net

carrier density variation, thickness, and thickness variation of the layer. Optionally specified items are

silicon source gas and growing method, transition width and flat zone (definitions for which are still

under development), and post-epi mechanical properties and visual inspection characteristics.

As noted above, the format, like other standards, is continually subject to revision as requirements

change. At present, SEMI is considering tighter tolerances on wafer dimensions, limits for surface metal

concentrations, specifications for surface microroughness, definitions for flat zone and transition width

of epitaxial layers, and ways of detecting surface contamination by particles of size 0. 1 /xm and smaller.

As these items are developed and approved, they will be incorporated into fiimre editions of the wafer

specifications and the specification form format.

— 17 —

Other Developments

ASTM Committee F-1 is developing new procedures to address both the statistical process control and

black-box measuring equipment issues as well as developing more precise test methods for wafer

geometrical and electrical parameters and new methods for surface metal contamination. It is also

beginning to develop consensus reference materials based on the results of interlaboratory testing for other

parameters such as flatness, sheet resistance, and net carrier density of epitaxial layers.

NIST is developing new SRM sets for resistivity determination which will have a precision better than

that of the previously issued sets, and SRMs for oxygen in silicon are expected to become available later

this year. Work is also continuing on a long standing program of developing microelectronic test

structures. These structures provide a basis for characterizing the results of a device fabrication process.

They are useful both in the design and manufacturing phases of device fabrication, and provide a much-

needed tool for evaluating the products of microelectronic foundries. The emergence of these foundry

operations is resulting in another split in the industry. Where once device houses did all the work from

growing the silicon to designing and processing the devices, the day of the "fabless" device house has

arrived. This effectively separates the device design and processing functions. As with the other

interfaces in the industry, good communications and stable relationships across this boundary are essential

for good results.

Thus, despite the limited resources and a slower rate of production than the heady days of the sixties,

considerable progress is being made in developing the metrological tools for silicon material and process

technologies required for fabrication of high density circuits with sub-micrometer design rules.

In-Process Metrology

These activities, while essential, all relate to supplier-user interfaces. However, today the need is also

for in-line control of wafer and device manufacturing processes. Significant reductions in cycle time and

direct labor hours together with increases in output and equipment uptime can be achieved by increased

use of in-line sensors and data links [17]. This topic has received attention from both SRC and

SEMATECH. Both have conducted numerous workshops and SRC has established a university center

on metrology.

To focus more attention on the issue of in-process metrology, ASTM and SEMI joined with JEIDA,

JESSI,"* NIST, SEMATECH, and SRC to conduct an international workshop in November 1992 [20].

This workshop was charged to formulate and publish a plan for the development of measurement tools,

techniques, and standards for qualifying process equipment and for controlling and improving unit

processes. This plan will provide guidance and motivation for metrological research, development, and

standardization activities in the industry.

Nearly 75 individuals representing manufacturers of materials, process equipment, metrology instruments,

and devices; universities; and government agencies participated in the workshop. There were seven

working groups which covered film deposition (including epitaxy), contamination, implantation, etching,

oxidation and diffusion (both furnace and rapid thermal processing), lithography, and materials. The

JESSI — Joint European Submicron Silicon Initiative.

— 18 —

most heavily represented category was metrology instrument manufacturers; otherwise, the representation

in most of the working groups was reasonably well balanced, except that process equipmentmanufacturers were not represented in several of the groups.

Each working group met three times. The first session was a brainstorming session in whichmeasurement issues and requirements were developed. In the second session, the groups prioritized the

requirements and, in the final session, developed a set of recommendations for specific metrology

development and standardization activities.

In carrying out their discussions, the groups employed the following proposed definitions:

in-situ metrology — metrology in which the sensor is embedded in the process tool,

monitoring directly and in real time the effects of the process on product or process

variables, thus providing a real-time measure of the result of process conditions with the

possibility of active control.

in-line metrology — metrology in which a metrology tool is integrated into the process

line with all of the product flowing through it, thus providing data, which can be used

for both statistical modeling and feedback/feed forward process control, for up to 100%of the product on the effects of the preceding processes on the product.

on-line metrology— metrology in which a metrology tool is attached to the process line

and automatically receives transferred product at a predetermined sampling rate (up to

100%), thus providing data, which can be used for both statistical modeling and

feedback-feed forward process control, for up to 100% of the product on the effects of

the preceding processes on the product.

off-line metrology— metrology in which product samples are brought to the metrology

tool which is located remotely from the process line, thus providing diagnostic product

data that are useful in understanding, modeling, or troubleshooting a process but do not

provide a mechanism for feedback/feed forward control of a process.

Although the report of the workshop is yet to be issued, a number of critical technical issues were

identified in the preliminary summaries of the working group deliberations which were presented at the

end of the workshop. These included both material characteristics (such as surface microroughness,

thickness and resistivity of very thin films, and surface metal concentrations) and measurement technology

(such as in-situ particle detection in process tools, including those introduced by the process itself; in-situ

surface temperature measurement; patterned wafer inspection and analysis; and applications of test

element groups, or microelectronic test patterns). There was much discussion of P/T ratios, the ratio of

the precision of a test procedure to the specification limit for the parameter being tested. As a rule of

thumb, the P/T ratio should be < 10%; for normal distributions, this implies that

USL-LSL

where o is the standard deviation of the measurement result, USL is the upper specification limit and LSL

is the lower specification limit. Thus, for this case, the standard deviation of the measurement tool must

be l/60th of the desired specification range, or less. Not many measurements today have this kind of

precision, even on a single-instrument basis.

— 19 —

Detailed recommendations on specific topics were made by each of the working groups. The following

items suggest the common threads of many of these recommendations. The report of the workshopshould be consulted for the detailed recommendations.

• Develop improved modes of communication between disciplines and between suppliers and

users.

• Improve accuracy of critical test measurements to support both material

specifications and process control requirements, or, alternatively, adjust specification

limits so that they conform with the P/T ratios of available measurement technology.

• Improve process models in order to improve the prediction capabilities of in-process

measurements.

• Develop a cost-of-ownership model for metrology tools.

• Develop a wide range of in-situ measurement techniques.

• Improve spatial resolution and sensitivity of measurement techniques.

• Develop systems for efficiently converting raw data into useful information.

Increased focus on in-process metrology and control will diminish the current emphasis on testing of

materials to a particular specification, but in no way diminishes the need for materials characterization.

Until process models are significantly improved, it will be virtually impossible to control most processes

solely with in-situ or in-line measurements of process parameters, such as temperature, flow rates, and

time. Nondestructive in-situ and in-line measurements on product samples (which may be as great as

100%) will continue to be required for the foreseeable future. If anything, the demands on

characterization techniques will become more severe in terms ofboth sensitivity and repeatability, because

assurance that the level of "out-of-spec" units is in the parts-per-million range requires great measurement

precision. First, as discussed above, P/T ratios must be improved. Secondly, the sensitivity of manymeasurements is not adequate, as noted in the example of surface metals contamination on silicon wafers.

Thus, considerable improvement inmost materials characterization measurements remains a requirement.

The problems and issues discussed above are not unique to silicon and silicon-based devices. Many of

them apply equally well to other materials and device families. The ability to control the device process

and to trace specific device structures back to the starting material and the various elements of the process

by which they were made is critical to the improvement of both device performance and manufacturing

efficiency. The importance of the ability to make this connection is greater, the more direct the

relationship between material characteristics and device performance.

Industry-WroE Planning

Also, in November 1992, the Semiconductor Industry Association (SIA) conducted a workshop to lay out

the technological roadmap that the silicon device and integrated circuit industry will follow into the next

century. The SIA Semiconductor Technology Workshop [21] involved 180 invited experts from all

aspects of silicon device development and manufacturing, from the materials and equipment industrial

infrastructure, and from universities and Government laboratories having related programs. The

participants were devised in advance into 1 1 working groups. Each working group addressed a carefully

defined subject area chosen so that no significant area of device design or manufacture was left

uncovered.

The workshop dealt specifically with mainstream silicon metal-oxide-semiconductor (MOS) integrated

circuit technology. It sought to identify the technical directions and accomplishments needed to make

— 20 —

succeeding generations of devices at 2-year intervals over the next 15 years. This was a heroic objective.

The outcome was a comprehensive and detailed plan which resulted from careful planning and substantial

pre-work activity by each participant. Nevertheless, the results were necessarily incomplete. For

example, equipment developments are for the most part implied rather than clearly defined. Similarly,

metrological needs must be inferred from the context. The plan will be fleshed out by independently

sponsored specialist workshops and maintained technically current through periodic reviews by the SIA.

Of particular relevance to the subject of this report is the fact that metrology was identified by the

workshop as one of the seven pervasive technology competencies essential to the successful execution of

the roadmap. The work of ASTM, SEMI, and NIST (as outlined in the foregoing sections) was

acknowledged. Strong recommendations were made regarding the necessary course of measurement

development, and a target budget for metrology (5% of die cost, exclusive of packaging, test, and design)

was established. SEMI has already revised its U.S. standards development program plans to reflect the

needs identified by the workshop. NIST is doing the same with its metrology R&D plans.

The importance of standardization of wafer diameter (also discussed earlier in this report) was reaffirmed

by many of the workshop working groups. Reference to Figure 3 shows that wafer diameters have

increased with each DRAM generation since the early eighties. The introduction of large volumes of

200-mm wafers was delayed past its expected date because of both technical and economic issues. It

appears that the transition to the next larger wafer diameter will be even more difficult than the transition

to 200-mm wafers has been. The workshop concluded that identification of the next wafer size was one

of seven key industry structure issues affecting the technology environment. Previous transitions have

been justified by balancing the larger number of available die per wafer against the larger cost of process

equipment and facilities; costs of development of the new process technologies has not so far been

significant compared to other process evolution costs. However, the workshop concluded that "for wafer

diameters of greater than 200 mm, wafer size conversion issues cannot be ignored," and recommends that

a working group be established to develop more detailed models to evaluate the economic benefits of

moving to larger wafer sizes so that the wafer size evolution can be predicted in a timely manner.

The occurrence of this workshop is evidence of the increasing maturing of the semiconductor industry.

This event defined the needs of a major U.S. industry to an extent never before achieved for any

industry, and in doing so, the industry has expressed its needs for measurements, materials, and

manufacmring tools far more clearly than at any previous time. As noted in the earlier sections,

measurement developments in past years have occurred with much less guidance from the device industry.

This clearer pathway and greater cooperation among the interested parties (another workshop

recommendation) will perhaps lead to significant improvements in the way measurement science and

technology are developed and delivered.

CONCLUSIONS AND RECOMMENDATIONS

In drawing the various threads of the silicon metrology development together, we find that there have

been a number of key factors which led to successful results. Many of these would appear to apply to

similar developments in connection with other materials.

Key Factors

It all began with freely shared basic research results which provided a solid foundation for future

development and growth. After some time, there was a widely recognized need for standardization of

— 21 —

test procedures to facilitate orderly exchange of materials in the marketplace; without this recognition,

the development of standards stagnates. Along with the recognition of the need, there was available an

organizational vehicle, ASTM Committee F-1, which was richly supported by major corporations and

NBS. The cornerstone of standard test method development in this organization was direct involvement

of both suppliers and users in experimental verification of the validity and precision of the test procedures

— to the ultimate benefit of the participants, document quality, and the industry at large.

The need for common ground and a willingness to share relevant information was amply demonstrated

by the failure of Committee F-1 to develop standards for phosphor materials. Also, the need for the

involvement of experts in the field was demonstrated by the difficulties encountered in the attempts to

develop standards for trace analytical methods. Although the presence of these items cannot ensure

success, their absence is certain to produce failure.

A turning point in the application of standardized test methodology for silicon material properties

occurred with the standardization of silicon wafer dimensions in the early seventies. The key point here

was the specification of a limited set of dimensional parameters coupled with considerable flexibility in

application-sensitive parameters. Not only did this increase awareness of the standardized test procedures

(even those for which parameter values were not standardized), it also provided for orderly growth as

wafer diameters increased, laid the foundations for the development and growth of the merchant

semiconductor equipment industry, and accelerated the internationalization of wafer standardization.

Although the demands on silicon metrology continued to increase as a result of decreases in feature size

and increases in circuit complexity, die area, and wafer diameter, the driving force in standards

development gradually shifted from the material itself to process equipment. In addition, the main burden

for developing improved, publicly available test methodology shifted from the major device and systems

houses to relatively small producers of measurement equipment. Test method standardization is

complicated by the widespread existence of black-box instruments which contain the essence of the

measurement technique imbedded within their software; often this software is inaccessible to the user.

To the need for increased sensitivity and reproducibility of both traditional and new measurements has

been added the need for increased in-process measurement capability in a total quality system

environment. These issues are currently posing severe challenges for the metrology and standards

communities in the silicon industry.

A good methodology for process diagnosis and control is an essential feature for applying test procedures

to the manufacturing environment. This was provided in the early days of integrated circuit manufacture

by the combination of MIL-STD-883 and the NASA Line Certification Program, but their effectiveness

was diluted by resistance to their application and the widespread use of waivers. Not until many years

later were statistical quality control procedures reintroduced into U.S. manufacturing environments, but

they are now being implemented throughout the semiconductor industry.

Finally, close cooperation between device and materials manufacmrers, first evidenced in the successftil

Japanese VLSI development in the early seventies and most recently recommended by the SIA

Semiconductor Technology Workshop, is necessary for efficient production of quality devices. Such

cooperation facilitates effective resolution of technological problems anmd provided definition for ftiture

materials and metrology developments.

— 22 —

Recommendations

Most of the developments we have discussed provide direct or indirect guidance for application to other

materials. For materials with small, specialized markets, the basics are most critical:

1 . Establish common ground and a willingness to share relevant information; develop

consensus standards for terminology, test methods for basic properties, and

specification formats to ensure that individually developed specifications are

meaningful and understood.

2. Involve experts in the development of standards; feed back the experience gained in

standards development to R&D groups for them to build further improvements.

3. Generate specifications with standardized values for parameters needed for orderly

flow of materials in manufacturing.

4. Assure that the measurement technology supports the necessary specifications with

regard to sensitivity and repeatability.

5. Apply statistical process control tools in a total quality system, using all available in-

process control metrology for both product and process parameters.

Finally, the issue of market size cannot be overlooked; for maximum effectiveness, no matter what else

is done, it is essential to take advantage of developments made by the silicon industry whenever they can

be applied, either unchanged or with modest modifications.

ACKNOWLEDGMENTS

Preparation of this report was supported in part by the National Institute for Standards and Technology under

Order No. 43NANB2 18856; this support is gratefully acknowledged. In addition, the important contributions

of many, many individuals, both at NIST (and its predecessor organization) and elsewhere, to the development

of silicon metrology and standards must be recognized. Without the dedication of these individuals, the

industry would not have the level of metrology support it currently enjoys.

REFERENCES

[1] W. Edwards Deming, Out of the Crisis (MIT Center for Advanced Engineering Study, Cambridge,

1982), page 278.

[2] CP. Marsden, "The Formation and Growth of Committee F-1," ASTM Standardization News 1 (4),

24-28, 54 (April 1973).

[3] W.M. Bullis and R.I. Scace, "Measurement Standards for Integrated Circuit Processing," Proc. IEEE

57, 1639-1646 (1969).

[4] Federal Standard 209A, "Clean Room and Work Station Requirements, Controlled Environment,"

August 10, 1966.

[5] W.M. Bullis, "Measurement Methods for the Semiconductor Device Industry—A Summary of NBSActivity," NBS Technical Note 511 (December 1969).

— 23 —

[6] J.W. Carlson, "Industry Standardization of Silicon Substrates," Solid State Technology 16 (9), 49-53

(September 1973).

[7] SEMI Specifications Ml, "Polished Monocrystalline Silicon Wafers," SEMI International Standards,

Materials Volume, published annually in April (SEMI, Mountain View, CA).

[8] G.E. Moore, Sr. and T. Law, "Worldwide Wafer Standardization—Fact or Fancy?" Solid State

Technology 18 (2), 40-43 (February 1975).

[9] W.A. Baylies, R.I. Scace, and F. Vieweg-Gutberlet, "International Standards for Semiconductor

Materials," Standardization News 11 (5), 21-23 (1983).

[10] W.M. BuUis, "Advancement of Reliability, Processing and Automation for Integrated Circuits with the

National Bureau of Standards (ARPA/IC/NBS)," Final Report, NBSIR 81-2224 (March 1981), 47

pages.

[11] ASTM Standard Test Method F 84, "Measuring Resistivity of Silicon Slices with a CoUinear Four-

Probe Array," Annual Book ofASTM Standards, Vol 10.05, published annually in November (ASTM,Philadelphia).

[12] CM. Osbum, H.L. Berkowitz, J.M. Heddleson, R.J. Hillard, R.G. Mazur, and P. Rai-Choudhury,

"Profiling of Ultra Shallow CMOS Junctions Using Spreading Resistance: A Comparison to SIMS,"

Proc. First Int. Workshop on Measurement and Characterization of Ultra-Shallow Doping Profiles in

Semiconductors, Research Triangle Park, NC, March 18-21, 1991, p. 433.

[13] A. Baghdadi, W.M. BuUis, M.C. Croarkin, Li Yue-zhen, R.I. Scace, R.W. Series, P. Stallhofer, and

M. Watanabe, "Interlaboratory Determination of the Calibration Factor for the Measurement of the

Interstitial Oxygen Content of Silicon by Infrared Absorption," J. Electrochem. Soc. 136, 2015-2024

(1989).

[14] M. Watanabe, "Technical Trends in Large Diameter Silicon Wafers," Solid State Technology 34 (3)

69-73; (4) 133-141 (1991). A full translation of the JEIDA report is being made available by SEMIunder the title, "Large-Diameter Silicon Wafer Trends."

[15] Proceedings ofthe Workshop on Silicon Materialsfor Mega-IC Applications , New Orleans, LA, January

30-31, 1991.

[16] SEMICON/West 91 Technical Program Proceedings, A SEMI/JEIDA Joint Technical Conference on

Silicon Materials for Advanced IC Applications, May 23, 1991.

[17] R.I. Scace, "Metrology for die Semiconductor Industry," NISTIR 4653 (September 1991), 41 pages.

[18] SEMI Format Ml 8, "Silicon Wafer Specification Form for Order Entry," SEMI International

Standards, Materials Volume, published annually in April (SEMI, Mountain View, CA).

[19] SEMI Specifications M2, "Silicon Epitaxial Wafers, " and Mil, "Silicon Epitaxial Wafers for Advanced

Applications," SEMI International Standards, Materials Volume, published annually in April (SEMI,

Mountain View, CA).

— 24 —

[20] International Workshop on Process Control Measurements for Advanced IC Manufacturing, Austin, TX,

November 9-10, 1992.

[21] SIA Semiconductor Technology reports: Vol. 1, Workshop Conclusions, and Vol. 2, Workshop

Working Group Reports (1993), Available from SIA, 4300 Stevens Creek Blvd., Ste. 271, San Jose,

CA 95129.

— 25 —

APPENDIX A

SEMI M18-93

FORMAT FOR SILICON WAFER SPECIFICATION FORM FOR ORDER ENTRY

Reprinted by NIST, with permission, from the Semiconductor

Equipment and Materials International Book of SEMI Standards

© 1990, 1993 Semiconductor Equipment and Materials International,

805 E. Middlefield Road, Mountain View, CA 94043

SEMI

SEMI M1 8-93

FORMAT FORSILICON WAFER SPECIFICATION FORM FOR ORDER ENTRY

1. Scope

1.1 This format provides a standard form for specifying

several classifications of silicon wafers, as follows:

1.1.1 Polished Silicon Wafers.

1.1.2 Epitaxial Silicon Wafers.

1.1.3 Epitaxial Silicon Wafers with Buried Layer.

1.2 The form is designed to be used in conjunction with

other SEMI specifications where details of the

dimensional, physical, electrical, and chemical properties

are defined or specified. However, the form includes

many items that are not currently included in the SEMIspecifications. Also, many items are not now commonly

specified, but may find increased importance in the future.

The intention is to provide for flexibility and expansion

of the technology.

1.3 This format provides for specifying and ordering

silicon wafers with varying levels of complexity. The

minimum level of completeness is shown in each case,

with additional options included to allow for

customization of the wafer to the specific processing

requirements of the user. If a particular item is not of

interest, no entry is required. The only required entries

are those marked as such for the purpose of minimal

specification (diameter, orientation, dopant, resistivity,

etc.).

1 .4 Values for many items are listed either in the polished

silicon wafer standards in SEMI M 1 or in the defect limits

tables in SEMI Ml, SEMI M2, or SEMI Mil. If these

standard specifications are sufficient for defining the

wafers, only a single entry is required in each section of

the form.

1.5 For referee purposes, U.S. Customary units shall be

used for wafers of 2- and 3-inch diameters, and SI (System

International, commonly called metric) units for l(X)-mm

and larger diameter wafers.

2. Applicable Documents

2.1 SEMI Specifications

SEMI Ml Specifications for Polished Monocrys-

talline Silicon Wafers

SEMI M2 Specifications for Silicon Epitaxial

Wafers

SEMI Mil Specifications for Silicon Epitaxial

Wafers for Advanced Applications

SEMI M12 Specification for Serial Alphanumeric

Marking of the Front Surface ofWafers

SEMIM13 Specifications for Alphanumeric

Marking of Silicon Wafers

SEMI M20 Specification for Establishing a Wafer

Coordinate System

2.2 Test methods referenced in Table 1 are cited in SEMIMl, SEMI M2, or SEMI Mil.

3. Definitions

3.1 The items listed in the form are referenced and

defined in various documents. A reference for most items

in Parts 2 and 3 of the form is listed in Table 1 . The entries

in the table are keyed to the form by item number. Be-

cause terms related to epitaxial wafers with buried layer

are defined in this standard and because no standardized

test methods for the buried layer parameters exist.

Table 1 does not include line items from Part 4 of the form.

3.2 General term used in this standard:

required (req. or req'd) - when applied to a parameter

listed in the order form, a user supplied value is necessary

to minimally define the material for manufacture.

3.3 Terms related to epitaxial wafers with buried layers:

alignment precision - pattern displacement in first mask

photolithography process.

NOTE 1 - Alignment precision is specified by maximumvalues of X and Y, the displacement of the center of the

pattern from a reference position defined in the wafer

specification in terms of the wafer coordinate system

defined in SEMI M20, and the maximum value of 9, the

angle between the x-axis of the pattern and the primary

orientation flat, see Figure 1.

pattern distortion ratio - Absolute magnitude of the

quotient of the (1) difference between the width of the

pattern on the substrate and the width of the pattern on the

substrate and the width of the pattern on the top surface

of the epitaxial layer and (2) the thickness of the epitaxial

layer.

201

SEMI Ml 8-93

©SEMI 1990, 1993

SEMI

Figure I

First mask pattern showing angular displacement, 9.

/epitaxicil layer

Si substrata

Figure 2

Cross sectional view of epitaxial layer showing the

pattern widths, a, at the epi surface, and b, at the

layer-substrate interface.

NOTE 2 - Pattern distortion ratio is specified by a

maximum value of Y, a dimensionless quantity given by

the quotient: Y = c/t,

where

c = \a-b\,

a = width of the pattern on the top surface of the

epitaxial layer, in ^m,

b = width of the pattern on the top of the substrate

wafer, in \im, and

/ = the thickness of the epitaxial layer, in \im.

See Figure 2. ^ > i x

pattern line width tolerance - allowed variation in pattern

line width on substrate surface.

NOTE 3 - Pattern line width is specified as i4 = X +Y,where X is the nominal value in mm and Y is the

tolerance in ^im. See Figure 3.

// Si02

Si substrate

Figures

Schematic diagram showing line width. A, of pattern on

substrate wafer.

pattern shift ratio - lateral distance between the centet

point of the pattern on the surface of the substrate and the

center point of the pattern on the surface of the epitaxial

layer divided by the epitaxial layer thickness.

NOTE 4 - Pattern shift ratio, d/t (see Figure 4), is specified

in terms of a nominal value, X, and a tolerance, ± Y, both

of which are dimensionless because both d and t are in

^im.

d

\V

AB

•pHaxial layar

Si substrate

Figure 4

Cross sectional view of epitaxial wafer showing the

pattern shift, d. Not to scale: AM=MB and CN=ND.

pattern step height - difference in vertical position of the

diffused (buried layer) surface and the original substrate

surface, after removal of oxide.

NOTE 5 - Pattern step height. A, is specified as a nominal

value, X, and a tolerance ± Y, both in nm. See Figure 5.

//

(//f//VJ Si substrate

A

buried layer

Figure 4

Cross sectional view of epitaxial substrate after oxide

removal but before deposition of the epitaxial layer

showing the pattern step height, A.

4. Use of the Form

4.1 If this form is used in conjunction with purchase

orders for silicon wafers. Parts 1 and 2 shall be completed.

For epitaxial wafers. Parts 1, 2, and 3 shall be completed.

For epitaxial wafers with buried layer. Parts 1, 2, 3, and 4

shall be completed.

4.2 In all cases, items listed as required must have a value

or choice indicated to minimally specify the material.

4.3 Certain required dimensional items may be specified

as a group according to the standard values presented in

the applicable SEMI specification, or they may be

specified individually.

SEMI Ml 8-93

©SEMI 1990, 1993 202

4.4 Visual inspection criteria may be specified as a group

according to the standard values listed in the applicable

SEMI specification, or they may be specified

individually.

4.5 For either dimensional values or visual inspection

criteria, the appropriate SEMI specification may be

marked and an optional line item (or items) marked as

well. In this case, the value marked on each individual

line item takes precedence over the standard value.

4.6 If the suggested form included in this format is not

reproduced and used as a fill-out form, the items and

responses must be adequately identified so that the

information and requirements are clear to all parties.

5. Test Methods

5.1 Measurements shall be made or certifiable to the

ASTM, JEIDA, JIS, or DIN standard test method as cited

in Table 1

.

5.2 When standard test methods from different

geographic regions are available, the default method shall

be the method in common usage for the region of the

purchaser of the wafer.

5.3 If several different standard test methods for an item

are commonly used within a region, a specific entry must

be made to identify which method of test is applicable.

5.4 If no standard test method for an item is available, the

test procedure must be specified.

6. Testing Level and Certification

6.1 Material ordered using this format shall be tested

and/or certified to the limits set forth in the individual

criteria.

6.2 The actual testing level, whether the test is performed

on all wafers, a sample of the wafers, special test pieces

representing all portions of the material, or a sample of

the test pieces, may be specified using this format.

6.3 In some cases, the material may be certified as

"capable of meeting" certain requirements. In this

context, the supplier is not required to perform the tests

outlined in this format. However, if the purchaser

performs the test and the material fails to meet the

requirement, the material may be subject to rejection.

203

SEMI Ml 8-93

©SEMI 1990, 1993

SEMISEMI SILICON WAFER SPECIHCATION FORM FOR ORDER ENTRY (Page 1)

PARTI. GENERAL INFORMATION

ITEMS INFORMATION1 Date:

Customer Name

Purchase Order Number

Line Number

Item Number

General Specification Number

Revision Level

Part Number/Revision

PART 2. POLISHED WAFERCUSTOMER: PART NUMBER: REVISION DATE:

RE

QITEM SPECIHCATION

TESTING LEVEL

WAFER TEST PIECE

100% SAMPLE 100% SAMPLE

1 . GENERAL CHARACTERISTICS

1.1 Growth Method ( jCz [ ]FZ [ ]MCz1.2 Crystal Orientation [](100) [Kill) [] ( )

1.3 Conductivity Type (Ip [ In

1.4 Dopant [ ) B [ ] Phos [ ] Sb [ ] As [ ] NTD1.5 Nominal Edge Exclusion

Distance for Fixed

Quality Area

6.10-6.14; 11 [ ]3mm [ )5 mm [ ] other mm7,8,12,13 [ )3mm [ ] 5 mm [ ] other mm(except edge defects)

2. ELECTRICAL CHARACTERISTICS

2.1 Resistivity Nominal[ ] ± Tolerance [ ) Q-cm

Measurement Position I ] Center Point ( ] Other:

2.2 Radial Resistivity Variation

(RRG)Not greater than [ ] %

Measurement Position ASTMF81 Fig. l-[ ]A, [ ]B, [ ]C, [ ]D; [ ] JEIDA

Calculation Method [ 1 ASTM F 81 [ ] JEIDA [ ] Other:

2.3 Resistivity Striations Not greater than ( ] %2.4 Minority Carrier Lifetime Greater than [ ] lis

Test Method ASTM F28 [ ] A, [ ]B; ASTM F391 [ ] A, [ ]B; [ ]other

3. CHEMICAL CHARACTERISTICS

3.1 Oxygen Concentration18 3

Nominal [ ] ± Tolerance [ ] x 10 cm or ppma

Calibration Factor [ lIOC-88 f ] old ASTM [ 1 Other

Measurement Method [ ]IR[

]SIMS [ ]GFA

3.2 Radial Oxygen Variation Not greater than [ ] %Measurement Position ASTM F 951 Plan ( ]A [ ]B [ ]B-1 [ ]C [ ]D [ )Other:_

Calculation Method ASTM F 951 Plan [ ]A [ ]B [ ]B-1 [ ]C [ ]D [ ]Other:_

3.3 Carbon Concentration Not greater than[ ] ppma

3.4 Metallic Content Not greater than [ ] ppma

3.5 Surface Organics Not greater than [ ] /cm^

4. STRUCTURAL CHARACTERISTICS

4.1 Dislocation Etch Pit

Density

[ ] Not greater than ( ] /cm^

4.2 Slip[ ] None [ 1 Other

4.3 Lineage ( ]Nonef ) Other

4.4 Twin [ ) None r ) Other

4.5 Swirl [ ] Not greater than ( ] % of wafer area

4.6 Shallow Pits Not greater than [ ] /cm^

4.7 Oxidation Induced

Sucking Faults (OSF)

Not greater than [ ] /cm

Test Cycle ASTM F 416 Cycle-[ ]1 ( ]2 ( ]3 [ 1 Other

4.8 Oxide Precipitatcs(BMD)

Interstitial Oxygen

Reduction (AO,)

Range: [ ) to [ ) /cm^

[ ] ppma

Test Cycle ASTM Fl 2391 1 A, f 1 B; f 1 Other:

SEMI Ml 8-93

©SEMI 1990, 1993 204

SEMISEMI SILICON WAFER SPECIRCATION FORM FOR ORDER ENTRY (Page 2)

PART 2. POLISHED WAFER (continued)

CUSTOMER- PART NUMBER: REVISION: DATE:

RE

QTTES SPECIFICATION

TESTING LEVEL

i WAFER TEST PIECE

100% SAMPLE 100% SAMPLE

5. WAFER PREPARATION CHARACTERISTICS

5.1 Wafer ID Marking [ [ None [ ] SEMI M12 [ ] SEMI MI3 [ ] Other:

5.2 Front Surface TTiin Film(s) AppI [ ] None [ ] Description:

5.3 Denuded Zone [ ] None f ] Description:

5.4 Extrinsic Gettering Treatment ( ] None [ ] Description:

5.5 Backseal [ ] None [ ] Description:

5.6 Annealing [ ] None [ ] Description:

6. MECHANICAL CHARACTERISTICS

THE ITEMS LISTED IN THIS SECTION MAY BE

6.01 [ ] Specified According to SEMI Ml.

OR SPECIFIED INDIVIDUALLY

6.1 Dianfieter' Nominal [ ] ± Tolerance [ ] mm6.2 Primary Flat Length/Diameter

Notch Dimensions'

fl.SF.MTMl. fl Other mm

6.3 Primary Flal/Notch Orientation' [] SEMI Ml. [] Other

6.4 Secondary Flat Length' f] SEMI Ml. n Other mm[ ] None

6.5 Secondary Flat Location' [ ] SEMI Ml. r 1 Other \ ] None

6.6 Edge Profile' [ ] SEMI Standard [ ]Other

6.7 Thickness Nominal ( ] ± Tolerance [ ) ;im

6.8 Thickness Variation (TTV)' (] SEMI Ml. r 1 Other urn

Measurement Position []F533 []F657 []H611 [) Other

6.9 Surface Orientation [ ] 0.00 ± 0.50 ( ]2.50 ± 0.50 ( ] 4.00 ± 0.50

[ ] Other

6.10 Bow'( ) SEMI Ml r 1 Other ^im

6.11 Warp' ( ] SEMI Mlf 1 Other |im

6.12 Sori f 1 urn

6.13 natness/Globai Acronym:^ [ ] [ ] [ ] [ ] Value ^m6.14 Flatness/Site Acronym:^

f ] ( ) [ ] ( ) Value jim

Site Size x mm Total Sites

% Usable Area f ] Include partial sites

7. FRONT SURFACE VISUAL INSPECTION CHARACTERISTICS

THE ITEMS LISTED IN THIS SECTION MAY BE

i

7.01 [ ] Specified According to SEMI Ml TABLE 1

OR SPECIRED INDIVIDUALLY

7.1 Scratches^[ 1 SEMI Ml Table 1 [ ] Other

7.2 Pits^[ ] SEMI Ml Table 1 [ ] Other

7.3 Haze^ [ ] SEMI Ml Table 1 [ ] Other

7.4 Light Point Defects

(Particulate Contamination)^

( ) SEMI Ml Table 1 [ ) Other

7.5 Contamination/Area^ [ ] SEMI Ml Table 1 [ ] Other

7.6 Edge Chips^ ( ] SEMI Ml Table 1 [ ] Other

7.7 Edge Cracks'* [ ] SEMI Ml Table 1 [ ] Other

7.8 Cracks, Crow's Feet^ ( ] SEMI Ml Table 1 [ ] Other

7.9 Craters*

( ] SEMI Ml Table 1 ( ] Other

7.10 Dimples'* [ 1 SEMI Ml Table 1 [ ] Other

7.11 Grooves'' ( ] SEMI Ml Table 1 [ ] Other

7.12 Mounds'* [ ] SEMI Ml Table 1 ( ] Other

7.13 Orange Peel' [ ] SEMI Ml Table 1 ( ) Other

7.14 Saw Marks' [ ] SEMI Ml Table 1 [ ] Other

7.15 Dopant Striation Rings [ 1

7.16 Stains f 1

SEMI Ml 8-93

205 ©SEMI 1990. 1993

SEMISEMI SILICON WAFER SPECIHCATION FORM FOR ORDER ENTRY (Page 3)

PART 2. POLISHED WAFER (continued)

CUSTOMER- PART NUMBER: REVISION: DATE:

RE

QITEM SPECIFICATION

TESTING LEVEL

WAFER TEST PIECE

100% 1 SAMPLE 100% SAMPLE8. BACK SURFACE VISUAL INSPECTION CHARACTERISTICS

THE ITEMS LISTED IN THIS SECTION MAY BE

8.01 [ ] Specified According to SEMI Ml Table 11

OR SPECIFIED INDIVIDUALLY

8.1 Edge Chips^ [ ] SEMI Ml Table 1 [ ] Other:i

8.2 Cracks, Crow's Feet^ [ 1 SEMI Ml Table 1 { ] Other:

8.3 Contamination/Area' [ ] SEMI Ml Table 1 [ ] Other:

8.4 Saw Marks' [ ] SEMI Ml Table 1 [ ] Other:

8.5 Stains [ ]

8.6 Roughness [ 1

8.7 Brightness [ 1

9. OTHER CHARACTERISTICS

PART 3. EPITAXIAL WAFER

CUSTOMER: PART NUMBER: REVISION: DATE:

R TESTING LEVELE ITEM SPECinCATION WAFER TEST PIECEQ 100% SAMPLE 100% SAMPLE

10. GENERAL EPITAXIAL WAFER AND LAYER CHARACTERISTICS

10.1 Conductivity Type/Structure (]n/n+ []p/p+[ ) Other:

10.2 Dopant []B []Phos [lAs

10.3 Silicon Source Gas [ ] SiH4 [ ] SiHzCb [ 1 SiHCb [ ] SiCU

10.4 Growth Method Reactor [ ] Pressure [ ]

10.5 Net Carrier Density

(Resistivity)

Nominal [ ) ± Tolerance ( ]

(check unit below)

Units(

]10' 'atoms/cm^[ ]n«cm

Test Method 1 ]HgC-V; ASTMF419[ JGated, ( JUngated,

[ ]Schottky; ( ]ASTM F374 (4-Pt); [ ] ASTMF525 (SRP):f ] Other

10.6 Net Carrier Density Variation []SEM1M2 [] Not greater than [ ]%Measurement Position [ ] SEMI M2 [ ] SEMI Ml 1 [ ] Other:

Calculation [ ] SEMI M2 [ ] SEMI Ml 1 [ ] Other:

10.7 Thickness Nominal ( ) ± Tolerance [ ] |im

Test Method []FT-IR [] ASTM F 672 (SRP)

10.8 Thickness Variation [ ) SEMI M2 ( ] Not greater than ( )%

Measurement Position [ ] SEMI M2 [ ] SEMI Mil [ ] Other:

Calculation [ ) SEMI M2 [ ] SEMI Mil [ ] Other:

10.9 Transition Width Not greater than [ ] nmMeasiurement Position

Calculation

10.10 Flat Zone Not less than [ ] ^m i

Measurement Position

Calculation 1

10.11 Phantom Layer [ ] None [ ) Not greater than [ ] ^m1

'

1

SEMI Ml 8-93

©SEMI 1990. 1993 206

SEMI SILICON WAFER SPECIRCATION FORM FOR ORDER ENTRY (Page 4)

PART 3. EPITAXIAL WAFER (continued)

CUSTOMER: PART NUMBER: REVISION: DATE:

RE

QITEM SPECinCATION

TESTING LEVEL

WAFER TEST PIECE

100% SAMPLE 100% SAMPLE

1 1 . MECHANICAL CHARACTERISTICS

11.1 Bow Not greater than [ ] |im

11.2 Warp Not greater than [ ] |im

11.3 Son Not greater than [ 1 jim

11.4 Flatness/Global Acronym^:[ ][ ][ ][ ] Value |im

U5 Flatness/Site Acronym^:[ ][ ][ ][ ] Value ^mSite Size x mm Total Sites

% Usable Area [ ] Include partial s/b sites.

12. FRONT SURFACE VISUAL INSPECTION CHARACTERISTICS

THE ITEMS LISTED IN THIS SECTION MAY BE

12.01 [ ] Specified According to: [ ] SEMI M2 Table 1 or [ ] SEMI Ml 1 Table 11

1

OR SPECIFIED INDIVIDUALLY

12.1 Stacking Faults* ( ] SEMI M( ) Table 1 [ ] Other:

12.2 Slip" [ ] SEMI M( ) Table 1 [ 1 Other:

12.3 Large Point Defects* [ ] SEMI M( ) Table 1 [ ] Other:

12.4 Total Point Defects* [ ] SEMI M( ) Table 1 [ ] Other:

12.5 Scratches [ ] SEMI M( ) Table 1 ( ] Other:

12.6 Dimples* [ ] SEMI M( ) Table 1 [ ] Other:

12.7 Orange Peel* [ ] SEMI M( ) Table 1 [ ] Other:

12.8 Cracks/Fractures* [ ] SEMI M( ) Table 1 [ ] Other:

12.9 Crow's Feet* ( ] SEMI M( ) Table 1 [ ] Other:

12.10 Edge Chips* [ ] SEMI M( ) Table 1 [ ] Other:

12.11 Edge Crown* [ ] SEMI M( ) Table 1 [ ] Other:

12.12 Haze* [ ] SEMI M( ) Table 1 [ ] Other:

12.13 Foreign Matter* [ ] SEMI M( ) Table 1 [ ] Other:

13. BACK SURFACE VISUAL INSPECTION CHARACTERISTICS

13.11Contamination'' ( ] SEMI M( ) Table 1 [ ] Other:

14. OTHER CHARACTERISTICS

1

1

1

207

SEMI Ml 8-93

©SEMI 1990, 1993

SEMI SILICON WAFER SPECIHCATION FORM FOR ORDER ENTRY (Page 5)

PART 3. EPITAXIAL WAFER WITH BURIED LAYER

CUSTOMER: PART NUMBER: REVISION: DATE:

RE

QITEM SPECIFICATION

TESTING LEVEL

WAFER TEST PIECE

100% SAMPLE 100% SAMPl

15. PHOTOLITHOGRAPHY CHARACTERISTICS

15.1 Mask ID [ ] Number: [ ] [ ] None

15.2 Alignment Precision X max ( ) nm, Y max [ ] |im, 0 max (]°

Reference Point (Wafer Coordinate System):

X [ ] nun, y [ ] mm15.3 Pattern Line Width Tolerance Tolerance ± [ ) nm16. BURIED LAYER CHARACTERISTICS

16.1 Dopant [ ]B.f ]Phos,[ ]S.[ ]As

16.2 Diffusion Depth (xj) Nominal [ ] ± Tolerance[ ] iim

16.3 Sheet Resistance Nominal [ ] ± Tolerance [ ] Q16.4 Pattern Step Height Nominal [ ] ± Tolerance [ ] nm

16.5 Defect Density (Before Epi) Not greater than [ J/cm^

17. BURIED LAYER PATTERN CHARACTERISTICS AFTER EPI

17.1 Pattern Shift Ratio Nominal [ ] ± Tolerance [ ]

17.2 Pattern Distortion Ratio Max. [ ]

18. OTHER CHARACTERISTICS

1

i

!

Individual value is not required if wafer is specified according to SEMI Ml. .

^ Flatness Acronyms are determined from the Flatness Decision Tree in SEMI Ml, Apftendix Al.

' Individual value is not required if wafer is specified according to SENfl Ml, Table 1.

* Individual value is not required if wafer is specified according to Table 1 of SEMI M2 or SEMI Mil.

SEMI Ml 8-93

©SEMI 1990. 1993 208

SEMI

Table 1

Industry Standard References

POLISHED WAFERITEM VALUE STANDARD REFERENCE'^ STANDARD TEST METHODS

IS IN

SEMIIS

REO'D"SEMI JEIDA ASTM JIS JEIDA DIN

1. GENERAL CHARACTERISTICS

1.1 Growth Method NO YES Ml Notes

1.2 Crystal Orientation NO YES F26 18 50433

1.3 Conductivity Type NO YES Ml Note 5 F42 H607 50432

1.4 Dopant NO YES Ml Note 5 50438/3

1.5 Nominal Edge Exclusion Distance for

Fixed Quality Area

NO NO Ml Fig 1

2. ELECTRICAL CHARACTERISTICS

2.1 1 Resistivity NO YES Ml Sec 9 F 84. F 673 H612 18 50431

2.2 1 Radial Resistivity Variation (RRG) NO NO Ml Sec 9 F81 18 50435

2.3 1 Resistivity Striations NO NO F 154, F 525

2.4 I Minority Carrier Lifetime NO NO F 28. F 391. F 1388 H604 50440/1

3. CHEMICAL CHARACTERISTICS

3.1 Oxygen Concentration NO NO F 120, F 1188,

F 1189, F 1366

50438/1

3.2 Radial Oxygen Variation NO NO F951

3.3 Carbon Concentration NO NO F 120, F 1391 50438/2

3.4 Metallic Content NO NO3.5 Surface Organics NO NO4. STRUCTURAL CHARACTERISTICS

4.1 Dislocation Etch Pit Density NO NO F47 H609 50434

4.2 i Slip NO NO F47, F416 H609 50434

4.3 1 Lineage NO NO F47 H609 50434

4.4 1 Twin NO NO F47 H 609 50434

4.5 Swiri NO NO F416 H 614 24

4.6 Shallow Pits NO NO F416, F 1049 H 614 24

4.7 Oxidation Induced Stacicing Faults

(OSF)

NO NO F416

4.8 Oxide Precipitates (BMD)Interstitial Oxygen Reduction (AOi)

NONO

NONO

]

!F1239

5. WAFER PREPARATION CHARACTERISTICS

5.1 Wafer ID Marking YES NO M12. M13

5.2 Front Surface Thin Film(s) Applied NO NO5.3 Denuded Zone NO NO5.4 Extrinsic Gettering Treatment NO NO5.5 Backseal NO NO5.6 Annealing NO NO6. MECHANICAL CHARACTERISTICS

6.1 Diameter YES YES^ Ml Sec 9 27 F613 50441/4

6.2 Primary Flat Length/Diameter

Notch Dimensions

YES YES*^ Ml 27 F 671, F 1152 50441/4

6.3 Primary Flat/Notch Orientation YES YES'^ Ml Sec 9 •>-T F847

6.4 Secondary Rat Length YES YES'^ Ml Sec 9 F671 50441/4

6.5 Secondary Flat Location YES YES*^ Ml Sec 9 F847

6.6 Edge Profile YES YES'^ Ml Sec 5 F928 50441/2

6.7 Thickness YES YES*^ Ml Sec 9 27 F533 H611 50441/1

6.8 Thickness Variation (TTV) YES NO Ml Sec 9 27 F 533, F 657 H 611 50441/1

6.9 Surface Orientation YES YES Ml Sec 9 27 F26 18 50433

6.10 Bow YES NO Ml Sec 9 F534 H611

6.11 Warp YES NO Ml Sec 9 F657, F 1390

6.12 Sori NO NO Ml Sec 9

6.13 Flatness NO NO Ml Sec 9 F775 DEF43 50441/3

209

SEMI Ml 8-93

©SEMI 1990, 1993

Table 1 (cont'd)

Industry Standard References

POLISHED WAFERITEM VALUE STANDARD REFERENCE'^ STANDARD TEST METHODS

IS IN

SEMIIS

reqd"SEMI JEIDA ASTM JIS JEIDA DIN

7. FRONT SURFACE VISUAL INSPECTION CHARACTERISTICS^

7.1 Scratches YES NO Ml Table 1 26 F 523 H 614

7.2 Pits YES NO Ml Table 1 26 F 523 H 614 24

7.3 Haze YES NO Ml Table 1 26 F 523 H 614 24

7.4 Light Scattering Defects

(Particulate Contamination)

YES NO Ml Table 1 26 F523 H 614 24

7.5 Contamination/Area YES NO Ml Table 1 26 F523 H 614 24

7.6 Edge Chips YES NO Ml Table 1 26 F523 H 614 24

7.7 Edge Cracks YES NO Ml Table 1 26 F523 H 614 24

7.8 Cracks, Crow's Feet YES NO Ml Table 1 26 F523 H 614 24

7.9 Craters YES NO Ml Table 1 26 F523 H614 24

7.10 Dimples YES NO Ml Table 1 26 F523 H614 24

7.11 Grooves YES NO Ml Table 1 26 F523 H614 24

7.12 Mounds YES NO Ml Table 1 26 F523 H614 24

7.13 Orange Peel YES NO Ml Table 1 26 F 523 H 614 24

7.14 Saw Marks YES NO Ml Table 1 26 F523 H 614 24

7.15 Dopant Striation Rings NO NO 26 F523 H614 24

7.16 Stains NO NO 26 F523 H614 24

8. BACK SURFACE VISUAL INSPECTION CHARACTERISTICS

8.1 Edge Chips YES NO Ml Table 1 F523

8.2 Cracks, Crow's Feet YES NO Ml Table 1 F523

8.3 Contamination/Area YES NO Ml Table 1 F523

8.4 Saw Maries YES NO Ml Table 1 F523

8.5 Stains NO NO8.6 Roughness NO NO8.7 Brightness NO NO9. OTHER CHARACTERISTICS

EPITAXIAL WAFERITEM VALUE STANDARD REFERENCE'^ STANDARD TEST METHODS

IS IN

M2IS IN

MilIS

REQ^SEMIM2

SEMIMil

JEIDA ASTM JIS JEIDA DIN

10. GENERAL EPITAXIAL LAYER CHARACTERISTICS

10.1 Conductivity Type/structure NO NO YES Sec 3 Sec 3

10.2 Dopant NO NO YES Sec 3 Sec 3

10.3 Silicon Source Gas NO NO NO Sec 3 Sec 3

10.4 Growth Method NO NO NO Sec 3 Sec 3

10.5 Net Carrier Density

(Resistivity)

NO NO YES Sec 3 Sec 3 F 1392, F 1393

(F 374, F 398,

F 525, F 723)

50438

(50444)

10.6 Net Carrier Density Variation

(Resistivity Variation)

YES NO YES Sec 3. 5 Sec 3,5

10.7 Thickness NO NO YES Sec 3 Sec 3 F95, F 110 50437

10.8 Thickness Variation YES NO YES Sec 3. 5 Sec 3. 5

10.9 Transition Width NO NO NO10.10 Flat Zone NO NO NO10.11 Phantom Layer NO NO NO

SEMI Ml 8-93

©SEMI 1990, 1993 210

SEMI

Table 1 (cont'd)

Industry Standard References

EPITAXIAL WAFERITEM VALUE STANDARD REFERENCE'^ STANDARD TEST METHODS

IS IN

M2IS IN

MilIS

REQ^SEMIM2

SEMIMil

JEIDA ASTM JIS JEIDA DIN

1 1 . MECHANICAL CHARACTERISTICS''

11.

1

Bow NO NO NO Ml Sec 9 Ml Sec 9 F534 H611

11.2 Warp NO NO NO Ml Sec 9 Ml Sec 9 F 657. F 1390

11.3 Son NO NO NO Ml Sec 9 Ml Sec 9

11.4 Flatness NO NO NO Ml Sec 9 Ml Sec 9 DEF43 50441/3

12. EPITAXIAL LAYER VISUAL INSPECTION CHARACTERISTICS'*

12.1 Stacking Faults YES YES NO Table 1 Table 1 F522

12.2 Slip YES YES NO Table 1 Table 1 F 47, F 80,

F52350434

12.3 Large Point Effects YES YES NO Table 1 Table 1 F 523. F 815

12.4 Total Point Defects YES YES NO Table 1 Table 1 F 523, F 815

12.5 Scratches YES YES NO Table 1 Table 1 F523

12.6 Dimples YES YES NO Table 1 Table 1 F523

12.7 Orange Peel YES YES NO Table 1 Table 1 F523

12.8 Cracks/Fractures YES YES NO Table 1 Table 1 F523

12.9 i Crow's Feet YES YES NO Table 1 Table 1 F523

12.10 Edge Chips YES YES NO Table 1 Table 1 F523

12.11 Edge Crown | YES YES NO Table 1 Table 1

12.12 Haze YES ! YES NO Table 1 Table 1 F523

12.13 Foreign Matter ! YES ! YES ! NO Table 1 Table 1 F523

13. BACK SURFACE VISUAL INSPECTION CHARACTERISTICS

13.1 Contamination ' YES i YES 1 NO Table 1 | Table 1 F5231

14. OTHER CHARACTERISTICS

^ Item is deFined or described and/or a test procedure applicable to the item is included in the cited reference.

Value is/is not required to minimally specify a wafer.

cIndividual value is not required if wafer is specified according to SEMI Ml.

''in today's technology, it is possible to inspect for some of these items using automated laser scanning systems; however, a standard

test procedure has yet to be developed. Application of automated inspection must be agreed upon between supplier and user.

NOTICE: SEMI makes no warranties or representations as to the suitability of the standards set forth herein for any

particular application. The determination of the suitability of the standard is solely the responsibility of the user. Users

are cautioned to refer to manufacturer's instructions, product labels, product data sheets, and other relevant literature

respecting any materials mentioned herein. These standards are subject to change without notice.

SEMI Ml 8-93

211 ©SEMI 1990, 1993

ly J. LJ J. Technical Publications

Periodical

Journal of Research of the National Institute of Standards and Technology— Reports NISTresearch and development in those disciplines of the physical and engineering sciences in whichthe Institute is active. These include physics, chemistry, engineering, mathematics, and computersciences. Papers cover a broad range of subjects, with major emphasis on measurementmethodology and the basic technology underlying standardization. Also included from time to time

are survey articles on topics closely related to the Institute's technical and scientific programs.Issued six times a year.

Nonperiodicals

Monographs — Major contributions to the technical literature on various subjects related to the

Institute's scientific and technical activities.

Handbooks — Recommended codes of engineering and industrial practice (including safety codes)developed in cooperation with interested industries, professional organizations, and regulatory

bodies.

Special Publications — Include proceedings of conferences sponsored by NIST, NIST annualreports, and other special publications appropriate to this grouping such as wall charts, pocketcards, and bibliographies.

Applied Mathematics Series — Mathematical tables, manuals, and studies of special interest to

physicists, engineers, chemists, biologists, mathematicians, computer programmers, and others

engaged in scientific and technical work.

National Standard Reference Data Series — Provides quantitative data on the physical and chemicalproperties of materials, compiled from the world's literature and critically evaluated. Developedunder a worldwide program coordinated by NIST under the authority of the National StandardData Act (Public Law 90-396). NOTE: The Journal of Physical and Chemical Reference Data(JPCRD) is published bimonthly for NIST by the American Chemical Society (ACS) and the

American Institute of Physics (AIP). Subscriptions, reprints, and supplements are available fromACS, 1155 Sixteenth St., NW, Washington, DC 20056.

Building Science Series — Disseminates technical information developed at the Institute on building

materials, components, systems, and whole structures. The series presents research results, test

methods, and performance criteria related to the structural and environmental functions and the

durability and safety characteristics of building elements and systems.

Technical Notes — Studies or reports which are complete in themselves but restrictive in their

treatment of a subject. Analogous to monographs but not so comprehensive in scope or definitive

in treatment of the subject area. Often serve as a vehicle for final reports of work performed at

NIST under the sponsorship of other government agencies.

Voluntary Product Standards — Developed under procedures published by the Department ofCommerce in Part 10, Title 15, of the Code of Federal Regulations. The standards establish

nationally recognized requirements for products, and provide all concerned interests with a basis

for common understanding of the characteristics of the products. NIST administers this programin support of the efforts of private-sector standardizing organizations.

Consumer Information Series — Practical information, based on NIST research and experience,covering areas of interest to the consumer. Easily understandable language and illustrations

provide useful background knowledge for shopping in today's technological marketplace.Order the above NIST publications from: Superintendent of Documents, Government Printing Office,

Washington, DC 20402.

Order the following NIST publications— FIPS and NISTIRs—from the National Technical InformationService, Springfield, VA 22161.

Federal Information Processing Standards Publications (FIPS PUB) — Publications in this series

collectively constitute the Federal Information Processing Standards Register. The Register serves

as the official source of information in the Federal Government regarding standards issued byNIST pursuant to the Federal Property and Administrative Services Act of 1949 as amended.Public Law 89-306 (79 Stat. 1127), and as implemented by Executive Order 11717 (38 FR 12315,dated May 11, 1973) and Part 6 of Title 15 CFR (Code of Federal Regulations).

NIST Interagency Reports (NISTIR)—A special series of interim or final reports on workperformed by NIST for outside sponsors (both government and non-government). In general,

initial distribution is handled by the sponsor; public distribution is by the National TechnicalInformation Service, Springfield, VA 22161, in paper copy or microfiche form.

Vu

esoU

o"oc

u

ac

iS OS

— oO (Nl

W £ 3Q< " J3*» i5

C U. O J

c .t

3 ^

St: .a «

O CLi